Archive for 23w5106
Arrowleaf Cellars [pinot noir]
Posted in Statistics with tags 23w5106, Arrowleaf Cellars, British Columbia, Canada, Canadian wines, Kelowna, Okanagan Valley, Okanagan vineyards, Pacific North West, pinot noir, wildfire on October 20, 2023 by xi'an
Arrowleaf pinot noir
Posted in Mountains, pictures, Travel, Wines with tags 23w5106, British Columbia, Canada, Kelowna, Lake Okanagan, Okanagan vineyards, pinot noir on September 20, 2023 by xi'an
approximate computation for exact statistical inference from differentially private data
Posted in Books, Mountains, pictures, Running, Statistics, Travel, University life with tags 23w5106, ABC algorithm, Approximate Bayesian computation, Bayesian inference, British Columbia, Canada, data privacy, differential privacy, EM algorithm, exact inference, Journal of Privacy and Confidentiality, Kelowna, Okanagan Valley, UBCO on September 10, 2023 by xi'an
“the employment of ABC for differentially private data serendipitously eradicates the “approximate” nature of the resulting posterior samples, which otherwise would be the case if the data were noise-free.”
In parallel or conjunction with the 23w5601 workshop, I was reading some privacy literature and came across this Exact inference with approximate computation for differentially private data via perturbation by Ruobin Gong that appeared in the Journal of Privacy and Confidentiality last year (2022). When differential privacy is implemented by perturbation, i.e. by replacing the private data with a randomised, usually Gaussian, version, the exact posterior distribution is a convolution which, if unavailable, can be approximated by a standard ABC step. Which, most interestingly does not impact the accuracy of the (public) posterior, i.e. it does not modify this posterior when the probability of acceptance in ABC is the density of the perturbation noise at the public data given the pseudo-data. Which follows from the 1984 introduction of the ABC idea. On the opposite, EM does not enjoy an exact version, as the E step must be (unbiasedly) approximated by a Monte Carlo representation that relies on the same ABC algorithm. Which usually makes the M step harder, although a Monte Carlo version of the gradient is also available.
Kelowna two weeks later
Posted in Mountains, pictures, Running, Travel with tags 23w5106, airbnb, BC, British Columbia, Canada, forest fire, Kelowna, Lake Okanagan, Okanagan Valley, wildfire on August 19, 2023 by xi'an
off to Banff, almost [23w5106]
Posted in Mountains, Statistics, Travel, University life, Wines with tags 23w5106, Canada, differential privacy, ERC, ERC Synergy Grant, Kelowna, Ocean, Okanagan Valley, University of British Columbia, workshop on July 30, 2023 by xi'an
I am almost off to a BIRS workshop on differential privacy taking place on the Okanagan UBC campus rather than at the original BIRS place in Banff, three hours (drive) West, on the Okanagan Lake. When I saw the announcement for the workshop on the BIRS webpage, I saw it as a perfect opportunity to jump into this central topic of our ERC Synergy grant and I contacted the organisers applying for a potential participation, which was successful! This is quite exciting, as the topic of privacy is, and also because I know none of the participants and hence the immersion into a new community should prove most fruitful. While not in my beloved Banff!, the location does not hurt either, with a nearby lake perfect for morning swims, and surrounding mountains for a hike or two. Before heading East, closer to Banff, for a vacation week.
